文献类型：期刊文献

中文题名：改进MSPCA的研究与应用

英文题名：A study of the improvement of MSPCA and its application

作者：张昱君[1];刘爱伦[1]

机构：[1]华东理工大学自动化研究所,上海200237

年份：2006

期号：5

起止页码：6

中文期刊名：工业仪表与自动化装置

外文期刊名：Industrial Instrumentation & Automation

收录：CSTPCD

语种：中文

中文关键词：主元分析;小波分析;多尺度主元分析;TE过程

外文关键词：principal component analysis; wavelet analysis; muhiscale PCA; Tennessee Eastman process

摘要：针对PCA方法单尺度建模的局限性,将小波分析和主元分析综合起来,引入了主元子空间之间的差别的概念,提出了新的检测工具改进MSPCA。与MSPCA方法相比,改进MSPCA方法降低了随机误差对测量数据的影响,提高了过程性能监视和故障诊断的准确性。仿真实例验证了该方法的有效性。
To improve the performance of PCA whose modeling is limited to a single scale, the paper rounds up the wavelet analysis and the principal component analysis（PCA） , and then introduces the definition of the difference between the principal subspaces into a multivariate statistical process monitoring tool as an improved multiscale PCA. Which can decrease the influence of random errors on the measured data and increase the reliability of the multivariate statistical model, thus enhancing the accuracy rate for process monitoring and fault diagnosis. Simulations verify the effectiveness of this improved MSPCA method.